2.1 Test organism
The Iran earthworm company delivered the Eisenia fetida. They were carefully transferred to the laboratory in moist soil. Under laboratory conditions, the earthworms were kept for 5 days in boxes containing the original medium (a mixture of sand, clay loam soil, and peat (1:1:1 v/v)) and cattle manure free of veterinarian products and maintained at room temperature (20±1ºC). The moisture content of the medium was adjusted to 50%. Soil water content was measured every week, and moisture was adjusted to 50% of the maximum water-holding capacity by adding distilled water. Healthy and clitellated adult earthworms (weight of 350–600 mg) with completely developed clitella were used in toxicity tests (OECD 2004).
2.2 Herbicides
Formulated metribuzin (Sencor, WP 70%) was obtained from Bayer, Persian AG, Tehran, Iran, halosulfuron (Sempra, WG 75%) from Nufarm Company and flumioxazin (Pledge, WP 50%) were from Sumitomo chemical company. The stock solutions and dilution series of metribuzin, halosulfuron were prepared in distilled water (1100 and 1630 mg/l) and for flumioxazin (1.7 mg/l) in acetone (OECD 2004).
The dilution series were made immediately before the application. A water control was for metribuzin and halosulfuron herbicides. A positive control (using the same amount of distilled water as added in each test substance assay) and a negative control (using the same amount of acetone for solvent control) were tested (OECD 2004).
2.3 Toxicity test methods
2.3.1 Filter paper test
An eight cm glass Petri dish was lined with a filter paper (Whatman filter paper no. 1) cut into 8×3 cm in diameter without overlapping. Metribuzin and halosulfuron were dissolved in distilled water and flumioxazin in acetone solvent and prepared herbicide concentrations. Two ml solutions were pipetted into the Petri dish. Control Petri dishes were also run in parallel with the herbicides. Distilled water was used as the control for metribuzin and halosulfuron and distilled water and acetone for flumioxazin. After the acetone was evaporated in an airing chamber, well-developed clitellate adult worms were randomly selected, washed, and dried. Then, they were exposed (one earthworm per Petri dish) to 2 ml of different concentrations of herbicides for 24 and 48 h. The earthworms were kept on wet filter paper for 24 h at 20±1º C in the dark to have the gut contents purged before the dose-response test, and they were washed and dried before being weighed and introduced in the test (Xue et al. 2009). The filter paper test had three replications. Every Petri dish contained ten adult earthworms. Petri-dishes were kept in the dark at 20±1ºC at 80-85% relative humidity and covered with a plastic lid with holes to allow for aeration and prevent the earthworms from escaping the Petri dishes. The experiments were performed for 48 h treatment period, and the number of dead earthworms of each treatment was recorded of each Petri dish. The concentrations of metribuzin, halosulfuron, and flumioxazin applied singly were shown in Table 1. A Preliminary range-finding test to find an optimal dose range that caused 0-100% mortality was done for the single herbicides before the mixture experiments.
2.3.2 Soil toxicity test
An artificial soil consisted of 10% ground sphagnum peat (<0.5 mm), 20% kaolin clay (>45% kaolinite), 70% quartz sand (<0.2 mm), and a small amount of calcium carbonate to adjust soil pH at 6.0 ± 0.5 (OECD 2004). The water content was adjusted to 35%. Stock solution of metribuzin and halosulfuron in water were prepared. This stock solution was used to spike soil with the highest test concentration and further diluted for spiking soil with the lower concentrations tested. The desired amount of herbicides was dissolved in 10 mL distilled water or acetone and mixed into a small quantity of fine quartz sand for each concentration. The sand was mixed for at least 1 h to evaporate the acetone and then mixed thoroughly with the pre-moistened artificial soil in a household mixer. All soils were thoroughly mixed to achieve a homogenous distribution of the herbicides in the soil. The addition of distilled water adjusted the final moisture contents of artificial soil. About 0.650 kg of artificial soil (equivalent to 0.5 kg dry artificial soil) was added into one L glass jar for herbicides. Three glass jars, each containing ten adult earthworms, were used for each concentration. Positive and negative controls were prepared with 10 mL distilled water or acetone and no herbicide to ensure if the positive and negative control were not significantly different from each other. The earthworms were removed from the culture, rinsed with water, and placed on damp filter paper in the dark at 20±1°C, and the content of their guts was emptied 24 h before herbicide exposure (Xue et al. 2009). The earthworms were fed 5 g cattle manure. Test jars were weighed at the start, so water loss could be monitored weekly and replenished with deionized water if needed. All earthworms were weighted before test. The jars were covered with perforated plastic film to allow the exchange of air and kept at room temperature 20± 1º C with 80–85% relative humidity under 400–800 lux of constant light. The survival was recorded on 1, 7, and 14 days after the treatment. The concentrations of metribuzin, halosulfuron, and flumioxazin applied singly were shown in Table 1.
The earthworms were preconditioned for 24 h under the same conditions described above in the untreated soil before the dose-response test. The ten adult earthworms were purged for 24h and were washed and dried before being weighed and introduced in the test. Two preliminary range-finding test determined the concentration range of the herbicides to find an optimal dose-range to caused 0-100% mortality.
2.3.3 Mixture toxicity
On the basis of the measured LC50 values of the individual herbicides (A and B), the relative potency (r) was determined for the herbicides applied alone. LC50s concentration (µg ai. cm-2 or mg ai. kg-1) reduced the live earthworm number by 50%.
Eq. 1 expresses the biological exchange rate between the herbicides, metribuzin, halosulfuron, and flumioxazin within the filter paper and the soil toxicity test. The LC50’s were derived from the dose-response regression model (Eq. 2) based on equation 1; the fixed mixture ratios were determined to ensure evenly distributed LC50 values for mixtures along the CA isobole (Gessner 1995). The mixture ratios were (100:0), (10:90), (25:75), (50:50) and (0:100)% for the metribuzin:halosulfuron mixtures and for metribuzin:flumioxazin they were (100:0), (4:96), (12:88), (29:71) and (0:100)% on filter paper test. The dose-response mixture experiments were independently repeated twice. Like the filter paper test, the mixtures ratios for the soil test were based upon the relative potencies of the individual herbicides. The mixtures were for metribuzin:halosulfuron (100:0), (72:28), (46:54), (22:88), (0:100)% and for metribuzin:flumioxazin they were )100:0), (25:75), (10:90), (4:96), (0:100)%. A positive water control was tested for metribuzin:halosulfuron and two water and acetone control for metribuzin:flumioxazin for filter paper and artificial soil tests. In neither case were there significantly difference between the positive control water only and negative control water plus acetone.
2.4 Statistical analysis
The dose-response data were analyzed using the R program (Version 3.6.1). The lethality of earthworms in response to herbicides is classical in toxicology. The binomial response, dead or alive, was assessed at various times during the experiments. The log-logistic regression was used to assess the acute toxicity of metribuzin, halosulfuron and flumioxazin, and the metribuzin:halosulfuron and metribuzin:flumioxazin mixture ratios (Eq. 2). The add-on R package drc (Version 3.6.1) was used to fit the log-logistic curves, and the LC50 parameters for each curve of fixed ratio mixtures illustrated the deviation from the straight line isobole of the CA reference model (Ritz et al. 2015).
The non-linear regression analysis of the log-logistic model (Ritz et al. 2015) is seen below.
Where y is the binomial response, dead divided by total number of earthworms in a petri dish or artificial soil, x denoted herbicide concentration (µg ai. cm-2 or mg ai. kg-1) of any mixture ratios defined as the sum of the actual doses. It means we fitted a total of five dose response curves per experiment for the binary mixtures. The experimental design was a so-called ray design (Gessner 1995). LC50 is the concentration (µg ai. cm-2 or mg ai. kg-1) that kill 50% of the total number of earthworms, and b is the slope of the curve around LC50. The dose-response fitted reasonably well to the data (See test for lack of fit Table 2).
The mixtures of metribuzin:halosulfuron and metribuzin:flumioxazin reference model CA at the equivalent LC50 doses can be expressed by (Streibig et al. 1998):
Z1 and Z2 are the LC50 of herbicide 1 and 2 applied singly, and z1 and z2 are the LC50 of herbicide 1 and 2 in a mixture adjusted by the relative potency, r (Eq. 1) herbicide 1 and 2 is applied alone in Eq. 1.